Opened 13 years ago
Closed 8 years ago
#9440 closed enhancement (fixed)
document more environment variables
Reported by:  John Palmieri  Owned by:  Minh Van Nguyen 

Priority:  minor  Milestone:  sage6.4 
Component:  documentation  Keywords:  
Cc:  Leif Leonhardy  Merged in:  
Authors:  John Palmieri  Reviewers:  KarlDieter Crisman 
Report Upstream:  N/A  Work issues:  
Branch:  dbea30c (Commits, GitHub, GitLab)  Commit:  dbea30c4b9a705a80e7f3c81d2f2a6a739a79e9c 
Dependencies:  Stopgaps: 
Description
In ticket #8263, we missed these:
 SAGE_DOC_JSMATH  should force docs to be built using jsMath
 SAGE_DOCBUILD_OPTS  passed as an argument to
sage docbuild
in the main Sage makefile.
Change History (18)
comment:1 Changed 13 years ago by
Cc:  Leif Leonhardy added 

comment:2 followup: 7 Changed 12 years ago by
TODO:
 Add
PARI_EXTRA_OPTS
, which is passed to PARI'sConfigure
when building Sage / installing the PARI spkg.  Add
CPPFLAGS
,LDFLAGS
,CXXFLAG64
?,LDFLAG64
?LD
?  Document that (e.g.)
export CFLAGS=""
has not the same effect asunset CFLAGS
.  Add note on not putting a space between
j
andNUM
in the description ofMAKE
, because doing so has a "completely" different meaning...  Add warning to (setting)
SAGE_CHECK
, since this currently breaks most builds. This will (or should) be fixed in future releases though; there's afaik a ticket adding more options to the value of this variable...  Add a note on / reference to
sageenv
.
Also, the description of some variables is incomplete, not current (in its use in all spkgs, e.g. SAGE_DEBUG
), misleading, or even wrong (SAGE_TIMEOUT*
).
I think the various variations of SAGE_T[E]MP[[_]DIR]
are missing. (I wonder where e.g. TMP
and TMPDIR
are / may be used instead, or as a fallback / default.)
SHELL
might be relevant, too.
The *ITER*
variables for (parallel) doctesting are missing.
I'm currently not sure if we have a similar section in the Developer's Guide, but some variables like SAGE_ROOT
and SAGE_LOCAL
should be documented (there?), too.
I'm still pretty sure even more variables are missing... ;)
comment:3 Changed 9 years ago by
Milestone:  sage5.11 → sage5.12 

comment:4 Changed 9 years ago by
Milestone:  sage6.1 → sage6.2 

comment:5 Changed 9 years ago by
Milestone:  sage6.2 → sage6.3 

comment:6 Changed 8 years ago by
Milestone:  sage6.3 → sage6.4 

comment:7 followup: 8 Changed 8 years ago by
SAGE_DOC_JSMATH
SAGE_DOC_MATHJAX
is now in good shape.
 SAGE_DOCBUILD_OPTS
if it still exists
 Add
PARI_EXTRA_OPTS
, which is passed to PARI'sConfigure
when building Sage / installing the PARI spkg. Add
CPPFLAGS
,LDFLAGS
,CXXFLAG64
?,LDFLAG64
?LD
? Document that (e.g.)
export CFLAGS=""
has not the same effect asunset CFLAGS
. Add note on not putting a space between
j
andNUM
in the description ofMAKE
, because doing so has a "completely" different meaning... Add warning to (setting)
SAGE_CHECK
, since this currently breaks most builds. This will (or should) be fixed in future releases though; there's afaik a ticket adding more options to the value of this variable...
This is basically fine now
 Add a note on / reference to
sageenv
.Also, the description of some variables is incomplete, not current (in its use in all spkgs, e.g.
SAGE_DEBUG
), misleading, or even wrong (SAGE_TIMEOUT*
). I think the various variations ofSAGE_T[E]MP[[_]DIR]
are missing. (I wonder where e.g.TMP
andTMPDIR
are / may be used instead, or as a fallback / default.)SHELL
might be relevant, too. The*ITER*
variables for (parallel) doctesting are missing. I'm currently not sure if we have a similar section in the Developer's Guide, but some variables likeSAGE_ROOT
andSAGE_LOCAL
should be documented (there?), too.
There is a ticket for the latter to be better, and the former is now in good shape.
I'm still pretty sure even more variables are missing... ;)
Indubitably.
comment:8 Changed 8 years ago by
Replying to kcrisman:
 SAGE_DOCBUILD_OPTS
if it still exists
It is actually used for docbuilding in the toplevel Makefile
. So it ought to be documented.
comment:9 Changed 8 years ago by
Branch:  → u/jhpalmieri/9440DOCBUILD 

comment:10 Changed 8 years ago by
Commit:  → 13513fa8760c747b2cb039d53c8e282030f7a128 

Status:  new → needs_review 
New commits:
13513fa  Trac 9440: Document SAGE_DOCBUILD_OPTS

comment:11 followup: 14 Changed 8 years ago by
Nice! Though I misled you by the wiki formatting. I don't know the status of these, and at least some definitely aren't documented.
 Add
PARI_EXTRA_OPTS
, which is passed to PARI'sConfigure
when building Sage / installing the PARI spkg.
Needed?
 Add
CPPFLAGS
,LDFLAGS
,CXXFLAG64
?,LDFLAG64
?LD
?
Needed?
 Document that (e.g.)
export CFLAGS=""
has not the same effect asunset CFLAGS
.
Needed?
 Add note on not putting a space between
j
andNUM
in the description ofMAKE
, because doing so has a "completely" different meaning...
Needed?
 Add a note on / reference to
sageenv
.
Needed?
Also, the description of some variables is incomplete, not current (in its use in all spkgs, e.g.
SAGE_DEBUG
), misleading, or even wrong (SAGE_TIMEOUT*
).
Needed?
I think the various variations of
SAGE_T[E]MP[[_]DIR]
are missing. (I wonder where e.g.TMP
andTMPDIR
are / may be used instead, or as a fallback / default.)
Needed?
SHELL
might be relevant, too.
Needed?
The
*ITER*
variables for (parallel) doctesting are missing.
Needed?
comment:13 Changed 8 years ago by
Commit:  13513fa8760c747b2cb039d53c8e282030f7a128 → 79280e197ff0507b3c2b29ed23fe7aefaedc9aae 

comment:14 Changed 8 years ago by
Replying to kcrisman:
Nice! Though I misled you by the wiki formatting. I don't know the status of these, and at least some definitely aren't documented.
 Add
PARI_EXTRA_OPTS
, which is passed to PARI'sConfigure
when building Sage / installing the PARI spkg.Needed?
Doesn't seem to be used, but some other variables are used in Pari's spkginstall file. I've documented those (just copying from spkginstall, basically).
 Add
CPPFLAGS
,LDFLAGS
,CXXFLAG64
?,LDFLAG64
?LD
?Needed?
I added a bit about them.
 Document that (e.g.)
export CFLAGS=""
has not the same effect asunset CFLAGS
.Needed?
Same here.
 Add note on not putting a space between
j
andNUM
in the description ofMAKE
, because doing so has a "completely" different meaning...Needed?
I don't know, this isn't Sage specific. I think we can skip it.
 Add a note on / reference to
sageenv
.Needed?
Done.
Also, the description of some variables is incomplete, not current (in its use in all spkgs, e.g.
SAGE_DEBUG
), misleading, or even wrong (SAGE_TIMEOUT*
).Needed?
I noticed that the documented value for SAGE_TIMEOUT was wrong. Beyond that, I don't know what Leif meant by this. If he wants to expand on it at some point, he can open another ticket and clarify.
I think the various variations of
SAGE_T[E]MP[[_]DIR]
are missing. (I wonder where e.g.TMP
andTMPDIR
are / may be used instead, or as a fallback / default.)Needed?
I don't think there are such variables anymore. There is a variable SAGE_TMP used while running Sage, but it's not usercustomizable and it's not an environment variable. So I don't think we need to document it.
SHELL
might be relevant, too.Needed?
Why? Let's skip it. What does it have to do with Sage?
The
*ITER*
variables for (parallel) doctesting are missing.Needed?
I added something about them.
comment:15 Changed 8 years ago by
Commit:  79280e197ff0507b3c2b29ed23fe7aefaedc9aae → dbea30c4b9a705a80e7f3c81d2f2a6a739a79e9c 

Branch pushed to git repo; I updated commit sha1. New commits:
dbea30c  Trac 9440: change date in source.rst

comment:16 Changed 8 years ago by
Oh, for what it's worth, I set SAGE_TUNE_PARI
to yes and rebuilt pari. Without this set, it took 4 minutes. With it set, it took 2 hours. So the comment about being timeconsuming is accurate.
comment:17 Changed 8 years ago by
Authors:  → John Palmieri 

Reviewers:  → KarlDieter Crisman 
Status:  needs_review → positive_review 
comment:18 Changed 8 years ago by
Branch:  u/jhpalmieri/9440DOCBUILD → dbea30c4b9a705a80e7f3c81d2f2a6a739a79e9c 

Resolution:  → fixed 
Status:  positive_review → closed 
I'm quite sure there are more missing... ;)