Frank Küster wrote: > Lutz Lehmann <[EMAIL PROTECTED]> wrote: > > >> ...It is not very helpfull ... >> to compress pdf files... >> > There's still a significant space saving when compressing pdf with gzip > (and there's zxpdf). This has been discussed on Debian-devel recently, > but no consensus has been found so far. Or maybe, just nobody bothered > to sum the discussion up. > to .pdf.gz: On a test example (math text with some graphs) perhaps 20 % savings, it may matter.
to zxpdf: It will never be invoked in the current setup, since the compressed file gets extracted to temp and is shown from there. In total, I believe that the suggested changes will make texdoctk work again. What remains of this discussion is a question of efficiency. For instance, it would be sufficient to note only the filename and not the path in the database. Then the direct lookup of the files could be removed in favor for the extended search method. In the current configuration, the latter method will always be called for compressed files, so efficiency is gained by omitting the first method. The database file will be smaller as well. Regards, Lutz
