an error condition when it gives warnings, so instead we grep
for warnings or error messages to see if we need to print the
log.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@476 788f1e2b-df1e-0410-8736-df70ead52e1b
+ Exit if pandoc fails (second time through) -- no need to store the log for this.
+ Run pdflatex up to three times, if needed to resolve references. Also
run bibtex as needed.
+ Minor reformatting.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@469 788f1e2b-df1e-0410-8736-df70ead52e1b
+ Add a trailing ':' to TEXTINPUTS as per the instruction in TeX FAQ:
http://www.tex.ac.uk/cgi-bin/texfaq2html?label=graphicspath
In the lack of it, pdflatex silently fails, for example, with the
following command: 'TEXINPUTS=/tmp markdown2pdf'
+ Put the origdir at the front for the correct directory search order.
+ pdflatex didn't create log file on one occasion (the above command) that
made sed commands failed. Test the existence of log before filtering it.
+ A few non-essential changes.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@459 788f1e2b-df1e-0410-8736-df70ead52e1b
--ignore-args features. This allows a simpler, cleaner design.
Make use of TEXINPUTS environment variable to ensure that
pdflatex will find images and other sources in the working
directory from which markdown2pdf is called.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@456 788f1e2b-df1e-0410-8736-df70ead52e1b
+ Improve sed filter to extract the following error contexts:
1. From a line starting with ! to the next blank line.
2. From a line beginning "LaTeX Warning:" to the next blank line.
3. From a line beginning "Error:" to the next blank line, or EOF.
+ Improve the error message headers (perhaps needs a proof reading).
Prepend the wrapper name to the error headers for easy spotting.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@451 788f1e2b-df1e-0410-8736-df70ead52e1b
/dev/null!
+ Another problem is the sed filter which returns nothing with pdfeTeX
'3.141592-1.21a-2.2 (Web2C 7.5.4)' here. As the first cut towards
fixing, use a somewhat heuristic approach: try to build a short log by
matching against a magic error stamp, dump the whole log if the previous
attempt failed. Note that, there is still room to improve this code.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@450 788f1e2b-df1e-0410-8736-df70ead52e1b
with our existing build process. There's no harm in having it
be a template, even though there's nothing to fill it.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@400 788f1e2b-df1e-0410-8736-df70ead52e1b
Cygwin, due to Windows' lack of true symbolic links).
+ Modified the wrappers to use 'pandoc' instead of the symlinks.
+ Modified the Makefile to remove all references to the symlinks.
+ Removed code from Main.hs that made pandoc's behavior depend on the
name of the calling program.
+ Added code to Main.hs that sets default reader and writer based on
extensions of input and output filenames (if provided). (Thanks to
roktas for the idea.)
+ Modified README and man pages accordingly.
+ Removed WINDOWS-README target from Makefile. It is no longer needed
now that we don't have the symlinks.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@295 788f1e2b-df1e-0410-8736-df70ead52e1b
Summary of main changes:
+ Added -o/--output and -d/--debug options to pandoc.
+ Modified pandoc to behave differently depending on the name
of the program. For example, if the program name is 'html2latex',
the default reader will be html and the default writer latex.
+ Removed most of the old wrappers, replacing them with symlinks
to pandoc.
+ Rewrote markdown2pdf and created a new wrapper web2markdown,
with the functionality of the old html2markdown script. These
new scripts exploit pandoc's -d option to avoid having to do
complex command-line parsing.
+ Revised man pages and documentation appropriately.
git-svn-id: https://pandoc.googlecode.com/svn/trunk@279 788f1e2b-df1e-0410-8736-df70ead52e1b
