SCIENTIFIC-LINUX-DEVEL Archives

February 2008

SCIENTIFIC-LINUX-DEVEL@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jon Peatfield <[log in to unmask]>
Reply To:
Jon Peatfield <[log in to unmask]>
Date:
Fri, 8 Feb 2008 22:12:06 +0000
Content-Type:
TEXT/PLAIN
Parts/Attachments:
TEXT/PLAIN (107 lines)
On Fri, 8 Feb 2008, Troy Dawson wrote:

> Troy Dawson wrote:
>>  Jon Peatfield wrote:
>> >  http://www.damtp.cam.ac.uk/linux/sl/damtp/5x/SRPMS/R-2.6.2-1.SL51.src.rpm
>> > 
>> >  The re-build finished quicker than I expected or the typing of that last
>> >  message took longer (or both).
>>
>>  Hi Jon,
>>  I got one change for your spec file.
>>  Right now when you get the release version you do
>>
>>       %define VER %(rpmquery --qf '%{VERSION}' sl-release | sed 's/\\.//g')
>>
>>  I have changed it to
>>
>>       %define VER %(rpmquery --qf '%{VERSION}' sl-release | cut -d'.' -f1)
>>
>>  This gives us the major release number (SL4, SL5) instead of the minor
>>  release
>>  number (41, 51).
>>  I like this better for two reasons.
>>  First - The rpm should be able to run on any SL4 or SL5 machine, so the
>>  minor
>>  number isn't really needed.
>>  Second (and the real reason I changed it) - Later on you have
>>       %if %{VER} >= 5
>>  And 41 is greater than 5, so it doesn't build on SL 4, because it's
>>  wanting
>>  everything that is in SL5.
>>
>>  OK, one other change, but this is very minor.  For the distro is has SL in
>>  capitol letters, while RedHat and Centos are in lower case letters (rh,
>>  centos).
>>  This is a personal thing, but I like it better in lower case letters.
>>  Like I said, this change is minor and purely a personal preference, but if
>>  you
>>  want it in capitol letters, leave it.
>>
>>  Thank you for your work on this.
>>  Troy
>
> OK, one other thing, maybe you know what's wrong with this one.
> This is on a SL 4 machine
>
> It's failling at the very very last because
>
> cp: cannot stat `doc/manual/R-exts.pdf': No such file or directory
>
> I can change the spec file and comment out
>
> %doc doc/manual/R-exts.pdf
>
> But since it actually builds on SL 5, I'm wondering if you would know of a 
> better way, or where that file comes from.

I had exactly that problem earlier when using our local texinfo (which is 
'too old').  It seems that the calls to texi2dvi on some of the doc fail 
unless a fairly recentish tex is installed.

[We still mainly use tetex-2 for 'compatability' but installing SL5's 
tetex-3 on the build setup worked fine]

Actually the Makefile that gets put in doc/manuals/ contains the 
following:

...
TEXI2DVI = /usr/bin/texi2dvi --texinfo="@set UseExternalXrefs "
TEXI2PDF = $(TEXI2DVI) --pdf
...

then it gets used like:

.texi.dvi:
 	@if test "$(TEX)" = false; then \
 	  echo "ERROR: 'tex' needed but missing on your system."; \
 	  exit 1; \
 	fi
 	@if test "$(TEXI2DVI)" = false; then \
 	  echo "ERROR: 'texi2dvi' needed but missing on your system."; \
 	  exit 1; \
 	fi
 	TEXINPUTS="$(srcdir):$$TEXINPUTS" $(TEXI2DVI) $<
...
.texi.pdf:
 	@if test "$(PDFTEX)" = false; then \
 	  echo "ERROR: 'pdftex' needed but missing on your system."; \
 	  exit 1; \
 	fi
 	@if test "$(TEXI2DVI)" = false; then \
 	  echo "ERROR: 'texi2dvi' needed but missing on your system."; \
 	  exit 1; \
 	fi
 	TEXINPUTS="$(srcdir):$$TEXINPUTS" $(TEXI2PDF) $<
...

causing 'test' to complain/fail.

Anyway with my older tetex on the path each attempt to make a .pdf file 
generates a bunch of errors but most of them do still produce .pdf files, 
and some completely fail.  Oddly enough cd'ing to docs/manual/ and running 
the 'make pdf' again gives *different* errors so something horrid is going 
on...

> Troy

ATOM RSS1 RSS2