Re: file path problems with pdf output
Ralf Angeli <[email protected]>
| Newsgroups | gmane.emacs.auc-tex |
|---|---|
| Message-ID | <[email protected]> |
* Björn Pedersen (2005-01-22) writes: > Following scenario: > > testdir_top: > dirtest.tex [...] > testdir_top/subdir: > subdirtest.tex: [...] > In pdf-mode, when trying to view from the subdirtest.tex buffer, > it prompts for xpdf ../dirtest.pdf, but executes xpdf in testdir_top, > so xpdf cannot find the pdf file. Manually removing ../ in the minibuffer > works. Specifing an absoulte path for TeX-master works correctly, too. [...] > It thems, the the %o expansion should not use the path part, or expand > to the absolute path of the master file. The %o expander calls `TeX-view-output-file' which subsequently calls `TeX-active-master'. Your former suggestion can be achieved by wrapping the call to `TeX-active-master' in `expand-file-name'. The latter suggestion by extending the definition of `TeX-active-master' by a parameter for specifying if the output should include the directory or not. That parameter will get passed to `TeX-region-file' or `TeX-master-file' respectively. Currently I tend to the latter solution because the shorter file specifier doesn't clutter the minibuffer as much as a full path. In addition it would be in line with what the user gets in the DVI case (%d expander). The drawback is that it might look strange to get the bare file name when the command is invoked in a subdirectory not containing the output file. A general problem might be that I haven't really had the time to look under the hood and understand how the viewer process actually gets called. So I don't know if there might be other hidden bugs with handling of files in subdirectories which are not covered by the patch. That means there is a possibility that a change on a deeper level would be more appropriate and fix the bug with the viewer and other potential bugs. -- Ralf