File Changes for max
switch to commits view, user indexv294_base to v295_preview (2014-01-13 to 2014-01-20) v295
- src/hg/hgGene/alleles.c
- lines changed 1, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/hgMirror/hgMirror
- lines changed 336, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- lines changed 5, context: html, text, full: html, text
making sure that trackDb actually exists before changing it
- lines changed 2, context: html, text, full: html, text
forgot to always mirror the slowest table of all, trackDb.
- lines changed 15, context: html, text, full: html, text
not trying to parse rsync log anymore, just checking at-queue now
- src/hg/hgTracks/wigTrack.c
- lines changed 25, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/hgVai/hgVai.c
- lines changed 11, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/hgc/pubs.c
- lines changed 12, context: html, text, full: html, text
breaking long lines in fasta view of pubs tracks, adding better sequenceIDs (refs #12212)
- src/hg/htdocs/goldenPath/help/trackDb/README.txt
- lines changed 1, context: html, text, full: html, text
tiny change to readme file
- src/hg/htdocs/goldenPath/help/trackDb/trackDbDoc.html
- lines changed 18, context: html, text, full: html, text
adding intro to containers in trackDb doc, refs #12486
- lines changed 19, context: html, text, full: html, text
fiddling again with the trackDb container intro, adding examples
- src/hg/htdocs/goldenPath/help/trackDb/trackDbHub.html
- lines changed 13, context: html, text, full: html, text
adding intro to containers in trackDb doc, refs #12486
- lines changed 18, context: html, text, full: html, text
fiddling again with the trackDb container intro, adding examples
- src/hg/htdocs/goldenPath/help/trackDb/trackDbLibrary.shtml
- lines changed 4, context: html, text, full: html, text
changing wording after code review rm #12486
- lines changed 5, context: html, text, full: html, text
adding intro to containers in trackDb doc, refs #12486
- lines changed 2, context: html, text, full: html, text
fixing stray div, refs #12079
- src/hg/inc/hdb.h
- lines changed 4, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/inc/jksql.h
- lines changed 3, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/inc/wiggle.h
- lines changed 1, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/lib/haplotypes.c
- lines changed 5, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/lib/hdb.c
- lines changed 135, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/lib/hgFind.c
- lines changed 24, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/lib/jksql.c
- lines changed 335, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- lines changed 5, context: html, text, full: html, text
fixing null pointer bug in jksql introduced with my last commit
- lines changed 8, context: html, text, full: html, text
not adding failover connections to the connection pool, as they aremanaged by their "parent". This lead to invalid pointers when failover
was activated in the sqlCleanup step upon exit. Also fixing a debug
message.
- lines changed 3, context: html, text, full: html, text
adding additional null pointer checks, just in case
- lines changed 36, context: html, text, full: html, text
Changing handling of sqlConnect to allow a db that exists onlyon the failover server, not on the main one.
- lines changed 5, context: html, text, full: html, text
handling the show table status command with failover connections
- src/hg/lib/pubsArticle.as
- lines changed 17, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsBingArticle.as
- lines changed 17, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsBingArticle.sql
- lines changed 5, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsBingBlat.as
- lines changed 30, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- lines changed 3, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBingBlat.sql
- lines changed 13, context: html, text, full: html, text
fixing stray div, refs #12079
- lines changed 2, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBingBlatPsl.as
- lines changed 26, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- lines changed 1, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBingBlatPsl.sql
- lines changed 1, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBingSequenceAnnot.as
- lines changed 12, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- lines changed 1, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBingSequenceAnnot.sql
- lines changed 1, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBlat.as
- lines changed 19, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- lines changed 12, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBlat.sql
- lines changed 13, context: html, text, full: html, text
fixed some more typos in pubs as/sql files
- src/hg/lib/pubsBlatPsl.as
- lines changed 26, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsBlatPsl.sql
- lines changed 28, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsMarkerAnnot.as
- lines changed 12, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsMarkerAnnot.sql
- lines changed 1, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsMarkerBand.as
- lines changed 9, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsMarkerBand.sql
- lines changed 1, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsMarkerGene.as
- lines changed 9, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsMarkerGene.sql
- lines changed 1, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsMarkerSnp.as
- lines changed 9, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/pubsSequenceAnnot.as
- lines changed 12, context: html, text, full: html, text
adding .as files for publications track, ref #12212
- src/hg/lib/search.c
- lines changed 8, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/hg/lib/wigDataStream.c
- lines changed 11, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/inc/linefile.h
- lines changed 5, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/inc/udc.h
- lines changed 9, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/lib/common.c
- lines changed 2, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/lib/linefile.c
- lines changed 42, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- lines changed 2, context: html, text, full: html, text
fixing a null pointer bug introduced in earlier commit today
- src/lib/maf.c
- lines changed 1, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- src/lib/udc.c
- lines changed 11, context: html, text, full: html, text
library changes for the browser box: This changes mostly hdb and jksql,plus - to a smaller extent - various other places in the code that deal
with /gbdb/ files. The overall aim is to make it possible to have the
data remote at UCSC while having the CGIs on a machine far away. At up to
180msecs distance from UCSC (Europe,Japan), each query can get slow. So
I tried to reduce the number of queries sent to UCSC while allowing to
keep some mysql tables on localhost.
I changed four things:
- extend larry's table cache to include field names. The code uses
"describe" very often, which is slow from remote. With a table name
cache these queries can be handled locally. This is configured in
hg.conf
- mysql "failover" connections: a mysql connection can have a 2nd
connection that is used if a query fails, configured in hg.conf
(I didn't call it "remote" connections, because we use that term already
in the code)
- mysql lazy connects: don't connect a sqlConnection right away, but
only when needed. a mysql connect takes >500msecs from across the
atlantic.
- move gbdb: patch various places that use absolute "/gbdb/" pathnames
to go through a central function that can change the filename of
gbdb files to something else, as configured in hg.conf
Plus patch 1 or 2 lines for more speed + update the hgMirror script
- lines changed 1, context: html, text, full: html, text
fixing a null pointer bug introduced in earlier commit today
- src/oneShot/sqlToAs/makefile
- lines changed 7, context: html, text, full: html, text
a quick .sql to .as converter
- src/oneShot/sqlToAs/sqlToAs
- lines changed 98, context: html, text, full: html, text
a quick .sql to .as converter
switch to commits view, user index