Convert DOCM to PDF

When Microsoft introduced the ECMA-376 formats in 2007 it split what had been one file type into two. A `.docx` is forbidden to contain a VBA project; a `.docm` is the same ZIP-of-XML container with permission to carry one. The registry records exactly that distinction — `macros: false` for DOCX and `macros: true` for DOCM — and it is the whole reason the second extension was invented.

The split was made so that software could tell the difference without opening the file. Before it, every Word document was potentially a program, and every mail filter, upload gateway and antivirus scanner had to treat them all as suspect. Afterwards a policy can say "DOCX yes, DOCM no" and be enforced by looking at four characters. That is enormously convenient for administrators and it is why your attachment disappeared.

Word macros do a characteristic kind of work, and it is worth naming because it decides what the PDF will contain. Most of them assemble: a template that builds a contract from selected clauses, a report generator that pulls figures in and formats them, a letter system that fills the address block and the reference number, a form that shows and hides sections depending on answers. The macro is a construction step that runs before the document is a document.

That is different from a spreadsheet, where the code usually operates on data that is already sitting in the file. It has a direct consequence here: if the macro has already run and the result was saved, converting gives you the finished document. If it has not — if the file is the template rather than an instance of it — the PDF shows placeholders and empty sections, correctly, because that genuinely is what the file contains. Open it in Word, let it build, save, and convert what was saved.

The conversion hands the file to LibreOffice running headless and asks for a PDF through its export filter. That path does not execute a VBA project, and it could not usefully do so — LibreOffice does not run Word’s macro language natively. Each job also gets its own freshly created LibreOffice profile, which exists for a different reason but has the same effect: nothing persists from one conversion to the next.

So the PDF is a rendering of the last saved state, and that is a guarantee rather than a limitation. It means converting a `.docm` you were sent and do not trust is a safe way to read it: whatever the macro was for, it does not get a turn. It also means the PDF is only as complete as the document was when it was written to disk, which is the point made above and the single thing most likely to surprise somebody converting a template by mistake.

Sending a `.docm` puts the recipient through a sequence of decisions you would rather they did not have to make. Word opens a file from outside the organisation in Protected View; if they choose to enable editing, a second bar appears asking whether to enable content, which is the prompt that turns the macros on. Many organisations now block macros from the internet outright, so the bar says the file has been blocked and offers no way through.

None of that is about your document. It is a blanket response to a decade in which macro-bearing attachments were the main delivery route for malware, and it applies whether the macro is a mail-merge helper or something worse. A PDF removes the whole conversation: it opens in the reader, in the browser, on a phone, in a preview pane, and nobody is asked to make a security decision about your contract on your behalf.

A PDF is a fixed page, so anything that was still calculating in the Word document resolves to its current value and stays there. Dates that updated themselves, page counts, cross-references, a table of contents, sequence numbers on figures and tables, DOCPROPERTY fields: each is frozen at the moment of conversion. Content controls and structured document tags become the text they were displaying.

That is usually the point — a filed or published document should not change after it is filed — but it is worth one deliberate check first. Update the fields and the table of contents in Word immediately before converting, so that what is frozen is correct rather than merely current. If any field was being populated by the macro rather than by Word, this is the case where an unrun macro leaves a visible gap in the PDF, and it is easier to see now than after the document has been sent.

The document is rendered by LibreOffice rather than by Word, so the layout is reproduced rather than reproduced identically. In practice it holds up well, because the converter carries metric-compatible substitutes for Arial, Times New Roman, Courier New, Calibri and Cambria — faces with the same character widths as the originals, so line breaks fall in the same places and a page laid out to fit still fits.

The risk is concentrated in unusual fonts. A brand face or a downloaded font that the document names but does not embed will be substituted with something of different widths, and text can reflow enough to push a line onto the next page. Documents assembled by a macro are more exposed than most, because template authors tend to be fond of a house typeface. If the document has a fixed layout that matters — a certificate, a form, a contract with numbered clauses on specific pages — look at the PDF page by page rather than at the first page only.

If the recipient has to edit the document, converting to PDF solves the security problem and creates a usability one. The better move is to save the file as `.docx` in Word: the format cannot hold a VBA project, so saving to it strips the macros while leaving an editable document that no filter will quarantine. Word will warn you that the project will be lost, which is exactly what you want.

Choose between the two by what happens next. Reading, printing, filing, publishing, attaching to a record — PDF, because it is fixed and universal and nothing about it can change. Editing, reviewing with tracked changes, filling in, or continuing the work — DOCX. Sending the `.docm` itself is the option to avoid unless the recipient specifically needs the automation and is expecting it, in which case a shared drive is a better delivery route than email.

On our server, not in your browser. Rendering a Word document to PDF needs a real office engine, and this page says so plainly rather than implying that everything on this site runs on your own device. The file travels over an encrypted connection to a container running LibreOffice headless, and that container is configured with no outbound internet access of its own — which for a file whose whole subject is untrusted code is worth stating explicitly.

Each job gets its own scratch directory and its own LibreOffice profile, and the directory is deleted when the job finishes regardless of the outcome. Two `soffice` processes sharing a profile corrupt it, and the symptom is a conversion that never returns rather than an error, which is why the profile is per job. Anything still running after sixty seconds is terminated. The free tier accepts 25 MB per uploaded file.

Read the PDF before it goes anywhere, and read it looking for absence rather than for damage. The characteristic failure of this conversion is not a mangled table, it is a section that should have been filled and is not, because the thing that fills it never ran. Empty address blocks, a reference number still showing its placeholder, a clause list with nothing in it, a total that reads zero.

Then check the page count and the last page. Documents assembled from templates tend to have carefully placed signature blocks and page-numbered schedules, and a font substitution that moves one line can push a signature onto a page of its own. If the same template is converted regularly, checking one output carefully tells you what to expect from all of them — the failures come from how the template was built rather than from any individual document.