On Sun, 19 Dec 1999, brian j pardy wrote:
> Yep, it was my fault. Here's a patch. Not intentionally obfuscated,
> but diff decided to do it this way.
>
> The old way:
>
> if (dump_output_immediately) {
> if (LYCookieSaveFile != NULL) {
> [tilde expand LYCookieSaveFile]
> }
> } else {
> StrAllocCopy(LYCookieSaveFile, "/dev/null");
> }
> } else {
> if (LYCookieSaveFile == NULL) {
> StrAllocCopy(LYCookieSaveFile, LYCookieFile);
> }
> }
>
> Now it attempts to expand LYCookieSaveFile before it worries about what
> mode we're in.
>
> diff -ru 2.8.3dev.17/src/LYMain.c 2.8.3dev.17.bri/src/LYMain.c
De-obfuscated by keeping only surviving lines:
> @@ -1605,29 +1605,28 @@
> LYLoadCookies(LYCookieFile);
> }
>
> + /* tilde-expand LYCookieSaveFile */
> + if (LYCookieSaveFile != NULL) {
> + if (LYCookieSaveFile[0] == '~' && LYCookieSaveFile[1] == '/' &&
> + LYCookieSaveFile[2] != '\0') {
> + temp = NULL;
> + StrAllocCopy(temp, LYCookieSaveFile + 2);
> + StrAllocCopy(LYCookieSaveFile, wwwName(Home_Dir()));
> + LYAddPathSep(&LYCookieSaveFile);
> + StrAllocCat(LYCookieSaveFile, temp);
> + FREE(temp);
> + }
> + }
> +
> /*
> * In dump_output_immediately mode, LYCookieSaveFile defaults to
> * /dev/null, otherwise it defaults to LYCookieFile.
> */
> + if(dump_output_immediately) {
The following is now happening unconditionally for dump_output_immediately,
even if a COOKIE_SAVE_FILE was explicitly given, that is overridden here:
> StrAllocCopy(LYCookieSaveFile, "/dev/null");
Maybe I de-obfuscated wrongly. The /dev/null thing should be subject to
'if (LYCookieSaveFile == NULL)', just like the non-d_o_i alternative below.
> } else {
> + if (LYCookieSaveFile == NULL)
> StrAllocCopy(LYCookieSaveFile, LYCookieFile);
> }
> #endif
Klaus
