On May 19, 2012, at 4:33 PM, Guy Harris wrote:
> In any case, I've thought about it a bit more, and think it might make sense to have, in the File menu:
>
> Export Packet Dissections
> Text...
> PostScript...
> CSV...
> PSML...
> PDML...
> Export SSL Session Keys...
> Export Objects
> HTTP...
> DICOM...
> SMB...
>
> rather than just "Export" as a top-level menu item - i.e., replace the "Export" top-level menu item with the three items that are currently underneath it, and give the Export->File item a name that indicates that what it exports is the result of dissection.
>
> They all "export" in the sense that they write out something other than the raw packet data, but they're all *very* different operations (the first group writes out, in non-capture-file formats, the result of packet dissection, whether it's the summary, details, or both, possibly with the raw data dumped in hex/ASCII format for text and PostScript; the second group writes out session keys from the capture; the third group writes out objects transferred over the wire by various protocols).
>
> The items under "Export Packet Dissections" would have longer descriptions similar to what they have now (I'm too lazy to type those in).
OK, I've checked in a change to do that.
I'll also be splitting "Save As" into:
File -> Save As - *always* writes out *every* packet, and, if it succeeds, makes the newly-saved file the current file;
File -> Export Specified Packets (verb subject to discussion) - *can* write out every packet, but offers the choices that "Save As" currently offers for selecting packets, and *never* makes the newly-saved file the current file;
which may clear up one of the issues from bug 6640.
(I've already changed "Save" so that, if you've edited the comments in a non-temporary file, it just saves on top of the existing file - using "safe save", i.e. "write to a new file and then rename the new file on top of the old file", as that's required by the way we work, and is *also* what should be done, so that writing the file out doesn't permanently destroy data if it fails in the middle.)